I am looking for two crosswalk files that will give me the relationships between county FIPS codes and codes for 1% PUMA files for 2000-2011 PUMAs and PUMAs from 2012 and later.
It does not appear that the MCDC can help me with this - an analyst at the Missouri Data Center explained that their utlility only uses 5% PUMA files and only from 2014 and later. I appreciate any guidance.
The 2000-2010 PUMA crosswalk file located here will allow you to crosswalk between 2000 and 2010 PUMAs. As you will see, the PUMA10_Name lists counties included. You can also use this map to to see boundaries for 2000 and 2010 PUMAs as well as 2010 counties.
You may also find the forum response found here useful, as it links to files from the Census containing the relationship between 2010 PUMAs and 2010 counties.
Thank you so much for the references and the quick reply, however, I am not yet convinced that I have the documents I need.
I am using the yearly ACS survey, which according to the first paragraph, page 2 of this document (https://www2.census.gov/programs-surv…), uses the 1 percent PUMA sample.
The document (https://usa.ipums.org/usa/volii/2000P…) referenced in the link provided in the reply shares information for the 5 percent sample (as does the MCDC utility).
It is my understanding that the one percent and five percent samples have different PUMA boundaries, because the 5 percent sample allows for fewer individuals in each PUMA, so the crosswalk between 5 percent and county should be different than the crosswalk between 1 percent and county.
Am I misunderstading the differences between the 1 percent and 5 percent samples? I appreciate any further guidance!
There is a substantial difference between the 2000 Decennial 5% and 2000 Decennial 1% samples. Super-PUMAs, which are geographic areas with 400,000+ residents, are the lowest level of geography in 2000 Decennial 1% samples, whereas PUMAs, which are lower level geographies than super-PUMAs, are identified in the 2000 Decennial 5%. As noted here, each super-PUMA is comprised of two to six PUMAs.
ACS samples, which are 1% samples, use the same PUMA boundaries as the 2000 Decennial 5% sample from 2005 - 2011. For ACS samples from 2012 onward, 2010 PUMA boundaries are used.
Are you using IPUMS USA data for your analysis? If so, you may also want to look into the CPUMA0010 variable.
I hope this helps.
Thank you for this reply! The consPUMA variable is very interesting. I don’t necessarily need it for my current study, since am not following a single PUMA over time. At the moment, I only need to compare a set of yearly ACS variables with a set of yearly county-level characterisics (collected to the PUMA level of geography). However, I’ll keep the consPUMA in mind for future needs!
Your reply has really helped me interpret the language in the technical documents that I’ve been reading. However, I want to be certain that we are on the same page. If you will indulge my reiteration of your instructions:
I am using ACS yearly data (1% sample) from 2000-2016 and need to find a county crosswalk.
I should use the FIPS county crosswalk for the 5% sample from 2005-2011 which are the same as the FIPS county codes for the 1% sample 2000-2011.
I should then rely on the crosswalk that allows me to compare 2000-2011 PUMAs to PUMAs from 2012-present to move from the 5% 2000 PUMAs to the counties in the 2012 1% PUMAs.
These two steps will allow me to observe which counties exist within each PUMA for the yearly/1% ACS data from 2000-present. Correct?
Thank you so much!!
The steps you have proposed should get you the information you are looking for.
You may also want to consider using the IPUMS USA COUNTYFIPS variable. With data from 1950 onwards, counties are unavailable in public-use microdata, however we attempt to identify as many counties as possible. Part of this identification uses PUMAs. More specifically, counties were identifiable if: they were coterminous with a single identifiable SEA, county group, or PUMA, or they contained multiple identifiable SEAs, county groups, or PUMAs, none of which extended into other counties.
This spreadsheet provides a list of counties and their corresponding ICPSR and FIPS codes available in each year from 1950 onwards.
This is great. I really appreciate your help.